Shanghai Launches China's Largest AI Computing Platform
Shanghai has unveiled a 10 billion yuan annual computing voucher program to build the nation's largest computing power scheduling platform, addressing a critical bottleneck in China's AI development. Announced at the 2026 Shanghai Global Investment Promotion Conference on March 14, the initiative provides a "use first, pay later" and "instant eligibility" mechanism to help enterprises access scarce computing resources.
The platform will provide access to 140,000 P of heterogeneous computing power, making it the largest such infrastructure in China. Large enterprises have historically struggled to find suitable computing resources, while SMEs have faced prohibitively high costs. This initiative directly addresses both pain points by aggregating supply and subsidizing access.
Shanghai also launched the nation's first corpus operation platform with 10,000 TB of datasets spanning AI4S (AI for Science), industrial manufacturing, and embodied intelligence domains. The platform has already connected 100,000 domestic and international developers, creating a foundational data infrastructure for AI training and research.
---
Kimi Valued at $18 Billion After Record Fundraising
Moonshot AI's Kimi has achieved a staggering $18 billion valuation following a new $1 billion financing round, representing a fourfold increase in just three months. The rapid valuation surge marks the fastest any Chinese company has reached unicorn status, setting a new record for domestic LLM continuous financing.
The company has completed three consecutive financing rounds in less than three months, demonstrating unprecedented investor confidence in China's large language model sector. Kimi has emerged as the flagship Chinese competitor to OpenAI's GPT series and Anthropic's Claude, gaining significant market traction since its initial launch.
This funding velocity reflects the intense capital competition in China's AI infrastructure sector, where compute availability has become the primary constraint on model development. Companies with guaranteed compute access have a decisive advantage in training next-generation models.
---
Former Huawei Execs Target $70B Data Center Power Market
Matrix Power, an AI data center (AIDC) full-stack power supply solution provider founded by former Huawei North America executives, has completed a tens of millions RMB Series A financing round led by Star Charge. The company has secured orders worth nearly 100 million RMB and is advancing partnerships with Google, Meta, and domestic AI server manufacturers.
The startup develops high-efficiency, high-density power products covering 54VDC in-cabinet power to ±400V/800V HVDC and solid-state transformers. These products address the massive power demands of modern AI data centers, which consume unprecedented electricity as compute clusters scale.
The global data center power equipment market is projected to reach 700 billion RMB by 2028, representing a massive opportunity as AI infrastructure expands worldwide. Matrix Power positions itself at the intersection of AI compute growth and energy infrastructure, a critical enabler for continued AI scaling.
---
Why This Matters
These three developments reveal a complete AI ecosystem emerging in China: government creating foundational infrastructure, private companies scaling rapidly on that infrastructure, and specialized hardware suppliers addressing the energy constraints that come with massive compute deployment.
Shanghai's initiative directly tackles the compute access problem that has constrained Chinese AI startups. By subsidizing access and building a scheduling platform, the city aims to democratize AI development beyond well-funded giants. Meanwhile, Kimi's valuation surge proves that companies able to secure compute resources can achieve world-class market positions.
The Matrix Power investment highlights an often-overlooked dimension of the AI race: power infrastructure. As AI models require exponentially more compute, data centers face electricity supply challenges that require innovative power delivery solutions.
Together, these developments signal that China's AI infrastructure buildout is accelerating across all layers—from data and compute access to energy supply—positioning the country as a global competitor in the AI era.